Skip to main content
ThinkEx uses Google’s Gemini models as its primary AI provider. You configure which providers are available by setting API keys in your .env file. Users can then switch between available models in the ThinkEx UI for each chat session.

​
Google AI (required)

ThinkEx requires a Google AI API key to power the AI chat features in every workspace.
​
GOOGLE_GENERATIVE_AI_API_KEY
string
required
API key for Google’s Generative AI (Gemini). All AI chat in ThinkEx runs through this key.
Get your API key:
1

Go to Google AI Studio

Open aistudio.google.com and sign in with your Google account.
2

Create an API key

Click Create API key and copy the generated key.
3

Add it to your .env

GOOGLE_GENERATIVE_AI_API_KEY=AIza...
Without this key, AI chat will not work. This is the only required AI credential.

​
Optional: Mistral OCR

​
MISTRAL_API_KEY
string
API key for Mistral AI. When set, ThinkEx uses Mistral’s OCR API to extract text from image-based PDFs and images uploaded to your workspace. Without this key, ThinkEx cannot read text from scanned documents or images.
Get a Mistral API key from console.mistral.ai. 
MISTRAL_API_KEY=your-mistral-api-key

​
Optional: Firecrawl web scraping

​
FIRECRAWL_API_KEY
string
API key for Firecrawl. When set, ThinkEx uses Firecrawl to fetch richer content from web URLs — including full page text and structured data — giving the AI more to work with when you add a website to your workspace. Without this key, ThinkEx falls back to basic URL fetching.
FIRECRAWL_API_KEY=fc_...

​
Optional: FastAPI service

The FastAPI service is an external backend that extends ThinkEx with additional file processing capabilities: converting office documents, extracting markdown from documents, and analyzing audio and video files.
​
FASTAPI_BASE_URL
string
The base URL of your FastAPI service (e.g., https://your-fastapi-service.com). Both variables must be set together for the integration to work.
​
FASTAPI_API_KEY
string
Bearer token used to authenticate requests from ThinkEx to the FastAPI service. Sent as an Authorization: Bearer header on all server-to-server requests.
FASTAPI_BASE_URL=https://your-fastapi-service.com
FASTAPI_API_KEY=your-service-api-key
The FastAPI service is optional and self-managed. ThinkEx does not provide this service — you deploy and operate it separately.

​
Model switching

Once you’ve configured your API keys, users can switch between available AI models directly in the ThinkEx UI. Model selection is per chat session and does not require any additional configuration.